Fitting a Bif sound system and updated lighting to a new GBRF liveried Bachmann Class 66.
I'm now living with the consequences of a decision reached when I was feeling sorry for myself while under attack from a rather nasty virus a couple of weeks ago. I decided that a bit of retail therapy was required to back up the anti-biotics! and discovered that using Hornby magazine's code, I could get £10 off the recent Bachmann GBRF Class 66 special "Modern Railways" 66745.
Having received the model, I ordered up a Legomanbiffo Class 66 LokSound V4 (plus 20x40 speaker) from Charlie at DC Kits etc. and topped up my LED stocks with some more nano-LEDs from Digitrains.
First thing to check is performance of the loco under DC control "as received" in case of any show stopping mechanical issues, before I carve anything up and kill the warranty!
Straight out of the box: It runs very smoothly from my bench power supply on a short test track.
The only problem was a low level "rattle" noise when under way. It was more apparent in one direction than the other.
My last 66 also suffered from this problem and it was caused by one of the bogies clunking on the chassis at the pivot point.
It was cured by fitting a small rubber washer on the bogie fixing screw. Hopefully the same fix will sort this unit out too!
First Update Issue....... Where to place the speaker?
The Bachmann unit includes a pre-fitted reduced depth 20x40mm speaker enclosure, positioned on top of the chassis block, below the roof fan grill.
However, I prefer the sound quality of a fuel tank mounted speaker, pointing down towards the track, as long as an acceptable enclosure can be created. Second choice (if this is unrealistic) is to mount the speaker via a custom sealed aperture, fixed directly to the fan grill surround, without an enclosure. This technique was used successfully on my Zimo equipped 66.
An extensive dismantling exercise revealed that it should be possible to create a suitable fuel tank mounted speaker solution! So this is the route I have chosen.
The 34 gram weight installed in the fuel tank has to be partially removed, but there is still room for some additional "liquid gravity" ballast to be added in the battery boxes, where the original rear light disable switch is located. (DCC switching will make this switch redundant). I've ended up reducing the original weight within the fuel tank by only 20gm. As the chassis block, motor and flywheel assembly weighs approximately 340gm, I don't think a 20gm loss will prove at all significant to the adhesion weight.
Switch and part of original weight removed. Switch blanking plate added and speaker hole cut.
The speaker is glued and sealed in place using Glue 'n' Glaze (still wet in this photo).
Underside view showing the ESU 50334 speaker cone. (After the glue had set!)
Liquid gravity replaces the switch. An edge shelf is glued in place to hold the sealing mastic.
An extra screw fixing hole is added and new leads are soldered in place and protected with electrical tape.
Re-assembly and speaker enclosure sealing:
The under frame with plasticard filling pieces added to minimise the gaps to be filled by mastic.
The underside of the motor, with blanking plate on the left of the insulated central core.
1) The fuel tank is first bolted to the plastic under frame, with a continuous strip of black mastic, sealing the edge of the speaker enclosure to the under frame.
2) The plastic under frame is next screwed to the underside of the metal chassis block. The motor protrudes below the chassis block, into the speaker enclosure and critically, there are air gaps around the motor, which need to be filled. A thin plasticard blanking plate was made to prevent mastic entering a vulnerable Hole in the motor casing. Further plasticard strips were glued to the periphery of the hole in the under frame through which the motor protrudes, to minimise the gap between the motor and the under frame. When the chassis block was re-attached, small quantities of black mastic were used to seal the remaining gaps.
The circuit has been updated throughout the project to incorporate any changes found necessary.
Showing the new LED mounted adjacent to the front wheel, just above the track.
Sorting Out The Missing Running Lights:
1. Adding a night headlight:
Bachmann provide the mounting pads for a night headlight LED and tracking to a rear wire connection pad, which makes this a simple mod.
2. Adding the lower front marker lights. (On the real Class 66/9, these are dual function rear and marker lights).
The lighting PCB support moulding includes cavities between the LEDs and the Lenses, to prevent unwanted light leakage. The inside surfaces of the cavities are painted matt white and pure white chip LEDs are then mounted in the top of the rear light cavities. The leads for the new LEDs pass through small holes drilled between the top of the moulding and the cavities.
With 2 to 3mA of current running through the LEDs, they are bright enough for the reflected light to give the right kind of intensity when viewed from directly ahead of the loco.
Lighting PCB assembly and mounting moulding temporarily re-assembled to check the effect.
Main PCB "Adjustments" for the modified lighting:
A bit of careful track tracing with a multi-meter indicates how the original Bachmann circuit handles the lights. However, in order to accommodate additional night running headlights and new lower marker lights, some changes are needed. I've decided to make two track cuts (to isolate the rear lights negative tracks) and remove 6 resistors from the main PCB and to add two small plasticard mounted circuit assemblies on either side of the main PCB to carry the new resistors and isolating diodes. (Diagrammatic Circuit shown below.)
The parallel 470 ohm resistors in series with the decoder common positive are shorted with a link as usual. (Still wonder what they are there for!)
The original lighting wires will also be re-attached to the 5 tabs at each end of the main PCB.
The connections for the MOSFETs will be shown later.
Fitting the MOSFETs & other circuit tweaks:
The MOSFETs are tiny surface mounting devices in SOT-23 packages
I've mounted the N-channel devices (2N2007) by super-gluing them, inverted, with legs in the air, on top of the main pcb. When the glue bond has dried, thin wires can (with some care) be soldered to the device legs.
However, the P-Channel device (BSS84) is connected in place of Bachmann's original rear light disable switch, which was wired to two tabs on the main PCB. So for convenience, I've soldered the Drain and Source leads directly to the appropriate PCB tags, with the 10k "hold off" resistor also soldered to the source PCB tag.
Dimming down the rear lights:
The rear lights are very bright! A simple way to reduce the LED current was to add a 3k9 resistor in series with each of the positive rear light connections to the main PCB (blue wires at each end of the board) the photo below shows how this was done.
Re-connecting the cab light for the unmanned rear cab:
I would have preferred to make the cab light(s) switchable, but had insufficient function outputs available. As a compromise, I connected the original front cab light via a 2k7 series resistor to the night running headlight function wire (Aux1) (to give a cab glow effect at night). If it is wished to make a similar mod to the other cab light, a 2k7 resistor can be connected between the Aux 2 function output and the pcb pad indicated in the photo below.
When the upper body shell connectors are plugged in, the rear cab light will then come on when the locomotive is reversing with night running lights on.
Programming the Decoder:
Button control of the lights:
3 buttons will operate the lights as follows: Button 0 = Running lights on/off; Button 11 = Rear lights disabled; Button 14 = day or night operation day/night. Direction will be set by the loco direction setting.
Detailed CV Changes to achieve new function mapping:
All working fine!
|Active Button Sound Slot
Assignments (& Loco Lighting Controls):
Motor Control Settings
The Updated Running Lights:
Day running lights
Night running lights
|A Few Hardware Issues to
1) A gate-source resistor was required to hold the P-Channel MOSFET off, when not switched on by the Aux 4 driven N-Channel MOSFET. I chose a value of 10k.
2) The rear lights were a bit on the bright side, so a resistor was added between each of the common positive blue wires from the lighting PCBs and the main PCB. 3k9 gave a more believable intensity.
The circuit diagram has been updated to include these modifications.
3) The driver's end will receive a solid buffer beam dam, coupling hook and hoses, while the rear end will employ a modified slotted buffer beam dam, to enable the tension lock coupler to operate....... and Oh yes! mustn't forget to glue the etched nameplates in place above the numbers on the side of the loco!
Name plate glued on
|A Few Software Issues to
1) I think the diesel motor sounds are a little on the quiet side compared to most of the ancillary sounds, but as the motor sounds (in sound slot one) arrived with the maximum volume setting of 128, I had to back-off most of the other sound levels to obtain (what seemed to me) to be a better balance. The sound level changes are documented above.
2) The lighting control was re-programmed so that function button 0 activates the running lights. Button 14 defines day (off) or night (on). While button 11 can be used to disable the rear lights when pulling a train. Function button 17 activates the detonator flash LED beside the front wheel and an orange "instrument glow" LED is activated in the occupied driver's cab whenever night lighting is selected. The updates all went smoothly and the necessary CV changes (obtained using the comprehensive ESU manual) are documented above.
3) Motor control CV values were tweaked as a result of previous experience with LokSound V4 decoders when driving Bachmann twin flywheel 3 pole motors of this size. The changes are documented above. Key issues are an expanded low speed part of the curve replacing Bif's linear values and a tripling of Bif's CV55 value to 100, reflecting the fairly high motor system momentum and hence (I believe) reducing the chance of any medium speed hesitation effects.
Next comes operational testing on the PC controlled test track and then the customary video to demonstrate how the new loco behaves.
Bif's Class 66 sound project is not behaving quite as well I expected!
Issue 1) When running light engine at slow speed, the engine notches up to pull away, then drops back to idle as intended.......but if the speed exceeds a scale 14mph, the engine notches up again and maintains notch 1. 14mph is rather slow, but the loco cannot be run at a higher sustained speed if you want it to remain in idle!
A fix: Bif sound projects are compatible with the ESU manual notch up and down commands:
Choosing a couple of so far unassigned buttons: Notch up on F15 and notch down on F16.
For F15: Programming line=21 Set CV32=3 Set CVN to 2 so CV333=2
For F16: Programming line=22 Set CV32=3 Set CVN to 4 so CV349=4
Now Set the throttle for the desired light engine running speed and shortly after the loco gets underway, hold down button 16 until the loco engine sound drops down to idle. It will then remain at idle revs until stopped or manually moved up a notch or two, using button 15! (Manual notching overrides Bifs system until the loco is stopped and returned to idle revs using button 16.)
Issue 2. (The biggest problem!) When the throttle is set near max at start up, Bif's projects normally respond by swiftly ramping up through the notches to simulate starting with a heavy train on the hook. However on my Bif 66 I found that more often than not a 3/4 max throttle start produced a step up to notch 1 at a fairly leisurely rate, a drop back to idle, a return to notch 1 and then a much more rapid ramp up through the remaining notch levels in the usual way. I have found that dropping Bif's original CV3 setting of 229 to 199, significantly improves the chance of a normal Bif ramp-up, using a 3/4 power throttle setting. In fact largely without failure after a reasonable number of runs!
Using manual notching, it is possible to wind up the notch range before rolling to simulate the power required to move a heavy train, but this process is slower than Bif's recommended use of max throttle and the notch 1 power fallback problem occurs every time, making the progress up through the notch levels sound rather non-prototypical.
The ESU notch up button timing is tricky even when used on a loco without the fallback problem...... with the notch 1 glitch happening its not really a useable solution.
However, it looks as if I can achieve the required end result for now, by dropping CV3 a little.
A comparison with Bif's 37, 57 & 67:
Issue 1: the 66 notch 1 transition speed is about half that of the other locos.
Issue 2: the other locos don't exhibit the 66 notch 1 power drop back when using a mid power throttle setting to start.
I've emailed Bif to see if the 66 driving behaviour is as he would expect. He responded quickly indicating that the 66 has 5 power notches compared to the other locos which have 4. This probably accounts for the lower initial notch transition speed. However the power drop back problem remained a mystery. He promised to investigate the issue when he regained access to his test 66 that weekend, in case it was a software issue. If the SW is confirmed OK, then that points to a hardware problem, specific to my decoder, in which case he'll replace it with a known good one! Can't say fairer than that! Impressive Bif service! ...... Even better: Bif confirms that he has found and fixed a software glitch causing the power drop back effect and he's re-blown my decoder to eliminate the problem! .....all within a couple of weeks! Now that's more than impressive!
Does the latest Digitrains Zimo Sound project work better?
I've now got the new Zimo decoder in my other sound equipped 66 up and running........ and it has its own rather different driving control issues...... There are some aspects that offer better user control than is possible on the LokSound: e.g. controllable start delay and adaptive acceleration for a realistic gradual start and 3 sets of selectable transition speeds between notches.....although I wish at least one of these options had higher transition speeds! On the downside, ancillary sounds are a bit thin on the ground: No drivers door slam, no buffering up clang and no detonators.
Having implemented the Zimo adaptive acceleration & deceleration option, giving a smoother initial pull away during a standing start, I found that I could no longer achieve the fast ramp up through the notches required to simulate a heavy train start, where considerable power is needed from the diesel engine to get the train moving. In search of a better solution, I've got a bit more radical, and using the available CVs, I've converted the original F5 Zimo "Light engine" mode to a new "power start" mode. I managed to lock up the chip at one stage and had to do a general reset, but I've found a way around that problem and it now works well. Normal mode (F5 off) displays smooth slow acceleration and includes flexible light engine operation, using F6 to hold or restore idle. The new F5, "power start" mode, has a fast ramp up through the notch levels, to simulate a start with a very heavy load on the hook. I've had to manually set a suitable value for the start delay as around 13 seconds is needed for a really heavy train and circa 4 seconds for light engine movements, making a compromise value unrealistic. I've also arranged for button F5 to switch between two of the notch transition setting sound sets. (The lower notch transition points being more suitable for a heavy train.) I'll get a video together to show how it works, when I've done a bit more experimenting....
Having got the Zimo decoder working in (I think) a somewhat improved manner, the two decoder systems have roughly equivalent performance. As usual its down to the personal preferences of the user and currently I'm happy with both solutions! (Although a few more ancillary sounds on the Zimo would be appreciated.)
Between 66745 with a fuel tank mounted installation and 66731, with a roof mounted installation.
Both speakers are ESU 50334 20x40mm units. They are only rated at 1.5Watts, so the volume from both decoder types has to be limited to a maximum of around half the available output level to avoid any danger of overload damage.
The fuel tank mounting provides a sealed enclosure a little larger in volume than the standard 12mm deep ESU enclosure. The roof mounting provides a sealed mounting between the speaker and the roof fan vent surround. However the locomotive body provides the much larger, but imperfectly sealed enclosure behind the speaker.
The perceived volume level from the speakers is very similar. The low frequency response of the roof mounted speaker appears very slightly more extensive than that of the fuel tank mounting, but otherwise, the tonal quality of the sounds is very similar.
I am sufficiently happy with the audio quality from both installations not to feel the need to retro-fit either unit with the alternative approach!
Bif LokSound V4 on the left.........
........Digitrains Zimo on the right
|Supplier website links: